Sensitivity analysis for attributable effects in case2 studies

Area of Science:

• Epidemiology
• Biostatistics
• Causal Inference

Background:

• The case-case study design is used for treatment effect inference.
• It compares treatment in 'cases of concern' to other cases.
• A key interest is the attributable effect, estimating cases that would not occur without treatment.

Purpose of the Study:

• To introduce a sensitivity analysis framework for case-case studies.
• To assess the impact of assumption deviations on attributable effect inferences.
• To evaluate unmeasured confounding effects in case-case designs.

Main Methods:

• Developed a sensitivity analysis framework for case-case studies.
• Applied the framework to assess deviations from key assumptions.
• Included sensitivity analyses for unmeasured confounding.

Main Results:

• The study provides a method to scrutinize inferences in case-case studies.
• Sensitivity analyses reveal the impact of assumption violations.
• The methodology is demonstrated using a real-world dataset.

Conclusions:

• The proposed sensitivity analysis enhances the robustness of case-case study findings.
• It addresses limitations of standard assumptions in real-data applications.
• This approach is valuable for causal inference in observational studies.